Persistence of dormant cancer cells that have resisted chemotherapy is a major barrier to curing ovarian cancer. Our group has developed the first inducible model for tumor dormancy in ovarian cancer using human ovarian cancer xenografts and has discovered a potential link between tumor dormancy and autophagy. Prolonged autophagy can induce cell death, whereas transient autophagy facilitates cell survival. We have found that ARHI (DIRAS3) - an imprinted tumor suppressor gene encoding a 26 kD GTPase homologous to Ras but with an opposite function - is required for autophagy in normal and malignant ovarian epithelial cells. ARHI participates directly in autophagosome formation, upregulating Atg4 and co-localizing in the autophagic vesicle membrane with MAP-LC3-II. ARHI also inhibits signaling through Class I PI3 kinase, ultimately decreasing the activity of mTOR, a known stimulus for the induction of autophagy. ARHI is downregulated in >60% of ovarian cancers. Re-expression of ARHI in ovarian cancer cells blocks proliferation by upregulating p21 and p27 and induces autophagic cell death in culture. In xenografts, re-expression of ARHI also blocks growth and induces autophagy, but cancer cells survive for months, remaining dormant and growing promptly when ARHI levels are subsequently reduced. Interestingly, growth factors (IGF-1, VEGF) and cytokines (IL-8) that are found in the xenograft microenvironment can rescue cancer cells from autophagic death in culture, block development of autophagy and partially reverse changes that occur in cell signaling during the induction of autophagy. Inhibition of autophagy with chloroquine inhibits subsequent growth of dormant cancer cells, but the precise relationship of autophagy and tumor dormancy remains to be defined. Other molecular abnormalities that occur frequently in ovarian cancers can regulate autophagy, including the overexpression of Rab25, but the interaction between ARHI and Rab25 has not been explored. In Aim 1 we will define the role of ARHI in autophagic vesicle formation, testing the hypothesis that ARHI associates directly with nucleating autophagic vesicles through lipid-lipid interactions, inducing Atg4 through FoxO3a to cleave MAP-LC3, facilitating vesicle membrane elongation. In Aim 2 we will explore the interaction of ARHI and Rab25 in regulating autophagy in ovarian cancer cells, testing the hypothesis that ARHI induces autophagy by inhibiting PI3 kinase activation, upregulating TSC1/2 and downregulating pmTOR, whereas Rab25 blocks ARHI-induced autophagy by activating Akt directly and increasing pmTOR activity. We also will determine whether expression of ARHI is associated with prolonged disease free survival, while expression of Rab25 will be associated with decreased progression free survival in ovarian cancer. In Aim 3 we will determine requirements for ARHI-induced dormancy in xenografts, testing the hypotheses that: 1) dormancy depends critically on the induction of autophagy; 2) survival of dormant autophagic cells in vivo requires permissive levels of VEGF, IL-8 and IGF-1; and 3) the accumulation of high levels of VEGF, IL-8 IGF-1 render ovarian cancer cells resistant to cell cycle arrest, autophagy and tumor dormancy. Further, we will seek autophagic cells that express ARHI in second look surgical specimens from women with ovarian cancer. If blocking VEGF, IGF-1 and IL-8 eliminates autophagic dormant ovarian cancer cells in xenografts, early translation to clinical trials should be feasible. PUBLIC HEALTH RELEVANCE: The persistence of dormant cancer cells that have resisted chemotherapy is the major barrier to curing ovarian cancer. These studies will test a novel strategy to eliminate dormant ovarian cancer cells by blocking factors required for their survival. If successful, early translation to clinical trials should be feasible.